This year we celebrate that it is 80 years ago that Roman Jakobson wrote his ‘Child Language, Aphasia and General Sound Laws’ (published in 1941), introducing the distinctive feature as a useful tool in our understanding of how sounds are structured in human language and human language behaviour. In this talk, we review the arguments in favour of the distinctive feature, both those put forward by Jakobson and those adduced later, as well as some of the problems. I then move on to better understand the internal structure of features: what kind of formal object could a feature be beyond an attribute-value pair, and propose that simple lambda calculus can shed new light on some problems that have traditionally vexed feature theory.